Micheal, >Lots of people run into this problem. The av_seek_frame seems to not work very well for this sort of stuff, so >a few months ago someone posted a library they wrote for the exact sort of purpose you describe. > >http://sourceforge.net/projects/ffmpeg-fas > >I haven't used it myself, but I've been meaning to see if it could clean up the code I use for playback of video >files. It has a very nice clean API. > >And, even if their library doesn't meet your needs, you can look at their code to see what approach they use.
Thanks for that. Not just me being stupid, then. The ffmpeg-fas project has been moved to: http://github.com/lbrandy/ffmpeg-fas I'm just finishing my OCR code, but I'll have a look at the fas code later and see if I can use it or adapt it to my needs. A quick scan through seems to show that it does a "pre-pass" through the video file to build a lookup table, which, in my case, is not too disastrous as I already do a "pre-pass" to determine the position of my target string's bitmap anyway. Clive _______________________________________________ libav-user mailing list [email protected] https://lists.mplayerhq.hu/mailman/listinfo/libav-user
